Improving the rigidity of the machining system in the cutting zone of ultra-thin-walled parts is a challenging problem to ensure machining accuracy, and the use of ice support is an effective method. However, ice and workpieces constitute a completely new process system that generates a complex process response under milling forces/thermal loads, which poses a challenge for deformation prediction. We investigate the milling deformation of ultra-thin parts from the perspective of analyzing the geometric nonlinear large deformation problem of plates. The equilibrium equations between internal forces and milling forces/heat in thin plates are established based on Kármán's assumption, and the equations for controlling the machining deformation of thin-walled parts under ice support are established by introducing the Winkler-Pasternak elastic foundation model, which considers the interfacial shear, to characterize the inverse restraining effect of ice on the ultra-thin-walled parts. Among them, the analytical model of milling force/heat under ice support is established separately in this paper as the load input to the deformation control equation. The displacement function in the form of trigonometric series is introduced to establish the stress function expression coordinated with the displacement, and the coefficients to be determined in the stress function are obtained based on the relationship between stress and deformation of the thin film in the midface. The nonlinear relationship between input load and deformation was solved by using the Bubnov-Galyokin method. A finite element model for milling of thin-walled parts under ice support was established, and the validity of the analytical model was verified by combining with milling experiments. The influence laws of cutting parameters, workpiece structural parameters and freezing condition on machining deformation are analyzed by means of analysis and simulation. Milling experiments were conducted on typical thin-walled parts and the uniformity of workpiece thickness after machining was evaluated. This work provides a theoretical basis for the prediction of machining deformation of ultra-thin parts under ice support.